A group of alpha testers have made OpenAI Sora, a generative artificial intelligence model for video creation, publicly available. They claim to have done so in protest against the exploitation of artists’ labor. This was reported by TechCrunch.

The group has published a project on the artificial intelligence development platform Hugging Face, which appears to be linked to the Sora API but is not yet publicly available. Using its authentication tokens-presumably from the early access system-the group has created a free and public version of the service.

“ARTISTS ARE NOT YOUR UNPAID R&D [Research and Design]. We are not your: free bug testers, PR puppets, training data, validation tokens,” the authors of the model write.

Any user could create 10-second videos with a resolution of up to 1080p by entering a short text description. OpenAI reacted quickly and shut down the model just 3 hours after it appeared on the platform.

The group claims that OpenAI misrepresents Sora’s capabilities by limiting early access users. All Sora results allegedly have to be approved by OpenAI before publication. Only a handful of authors participating in the program can display their work created using this model.

“We are not against the use of AI technology as a tool for the arts (if we were, we probably wouldn’t have been invited to this program),” the group writes. “What we don’t agree with is how this artist program has been rolled out and how the tool is shaping up ahead of a possible public release. We are sharing this to the world in the hopesthat OpenAI becomes more open, more artist friendly and supports the arts beyond PR stunts.”

OpenAI responded to this event and stated that work on the model is still ongoing and that the company is “working to balance creativity with strong security measures for widespread use.”

A company spokesperson also said that artists “have no obligations” to OpenAI other than to use Sora “responsibly” and refrain from disclosing confidential information while Sora is still in development. However, they did not specify what “responsible” use means, nor did they say what kind of data OpenAI considers confidential.